The five consensus algorithms #1: Proof-of-work by Dr. Leemon Baird


some DLTS are based on proof of work proof of work is the idea of using a supercomputer a mining rig that’s expensive to burn lots of electricity which is expensive in order to solve a math problem that’s totally useless to humanity but the purpose is to slow you down so that at slower intervals a miner will solve the problem and be able to add a new block to the chain and so the blocks come in slowly now occasionally two people add a block at the same time which is a fork that’s a bad thing and then the community as a whole has to come to an agreement on which of these two do we increase and which one do we ignore and it takes a little while for them to decide which one to continue with in which one to basically chop off because it takes a while we need to make sure they don’t start 14 too fast if you fork too fast you wouldn’t be able to chop them off as fast as they’re appearing and that’s the reason for the proof of work the purpose of proof of work is to make you run more slowly so that the community will have time to deal with these Forks before the next fork happens of course if the whole point of it is to make us run more slowly the result is we run more slowly and we’re wasting resources and the resources are expensive which means that we will tend to have consolidation of miners down to the few people who can afford the equipment and consolidation of geography down to the places where electricity is cheap there are also security problems what if there is a firewall around a good portion of our people the firewall is run by maybe someone malicious who decides to turn off the firewall partitioning us could we end up with two different chains or maliciously manipulates the firewall turning it on and off in some disruptive way or letting some packets through and not others could we end up with different consensus on each side six confirmations on each side and then a double spin wouldn’t be caught or what about fairness what if you’re trying to do something like a stock market where the ordering of the transactions matters well with proof-of-work whatever minor puts the transactions in that block can put them in any order that they want and they can leave some out if they want and maybe you bribed them to leave out your rivals transactions so we don’t really have fairness of ordering or fairness of access or even fairness of timestamps the minor puts a timestamp on the block but maybe they’re lying or maybe their clock is wrong you don’t really have the whole community weighing in on what time the timestamp should be so we have some difficulties with firewall attacks we have some difficulties with fairness of access and ordering and of the timestamps and then we have the expense and inefficiency of doing the waste of time but it does work and it does allow us to have a single chain that whenever it Forks we quickly chop off the forks then then the chain goes back to being a chain and that is the proof-of-work deal to technology

, , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , ,

Post navigation

6 thoughts on “The five consensus algorithms #1: Proof-of-work by Dr. Leemon Baird

  1. Could someone please organize a discussion between Andreas Antonopolous and Leemon Baird, that would be absolutely fascinating! 🙂 (I'm not joking either)

  2. New consensus algorithm in town! Prism: https://www.youtube.com/watch?v=WBTz_Gr0IUk. It achieves optimal performance.

Leave a Reply

Your email address will not be published. Required fields are marked *